I know some of the later stuff (SIMD / AVX / SSE / etc) is Intel, but technically I think they could leave that off and still be successful assuming that would be all that's needed to not be infringing (which I don't know if that's the case or not).
And Intel has been making claims that they could start to litigate. https://www.theregister.co.uk/2017/06/09/intel_sends_arm_a_s...
IIRC the previous gen A10x in the iPad Pro beat some low-end macs, and usually the next gen non-x is a bit faster than the previous gen x... so, I'd expect the A11 to be faster than some low-end macs.
Could Apple be arranging 3rd party conversions in secret? Or, could the conversion be done seamlessly, at the compiler level?
Anyway, I think ARM macs have been on the cards for a few years now... in a way, the iPad Pro can be seen as a way of experimenting with Mac-level power, as a preparatory step.
Their hesitation might partly be due to a kind of identity crisis: if we can develop iOS apps on a Mac, and that Mac is ARM, why can't we develop on an iOS device itself? But Apple doesn't want open iOS devices.
I think it's more likely that one of the purposes of the beta was to determine when thermal throttling was ideal, given real-world workloads (which are hard to fine-tune without knowing exactly how people will use it).
It's a pity thermal throttling is necessary; but at least in a Mac form-factor, there would be more cooling options.
I did check the Geekbench scores, because I was curious about your speed claims... https://browser.geekbench.com/v4/cpu/search?utf8=&q=7360u and https://browser.geekbench.com/ios-benchmarks/ seem way too close to equal for me to believe. Maybe that's the way it is, but I'm assuming I'm misunderstanding the tests. The A11 couldn't possibly be that fast, could it?